The ESR dose values of two human tooth enamel fragments were measured in 10
degrees steps with an automated goniometer. ESR spectra were deconvoluted
using four Gaussian peaks. The central region of the spectra is interfered
with by at least one additional line which yields significantly lower dose
values than the central, apparently axial region. There is a clear relation
ship between dose and width of the deconvoluted axial peaks which is attrib
utable to incomplete separation between the axial and interfering peaks. A
simple model is presented that coherently explains all observations. We con
clude that deconvolution is necessary for the estimation of the most likely
dose value of enamel pieces. (C) 2000 Elsevier Science Ltd. All rights res
erved.